- Email: [email protected]
Tribological behavior of DLC coatings with functionally gradient interfaces
Surface and Coatings Technology 153 (2002) 178–183
Tribological behavior of DLC coatings with functionally gradient interfaces Y. Liu, E.I. Meletis* Materials Science and Engineering Program, Mechanical Engineering Department, Louisiana State University, Baton Rouge, LA 70803, USA Received 6 June 2001; accepted in revised form 14 November 2001
Abstract Initial evidence showed that soft substrate materials (such as Ti–6Al–4V alloy) might not be able to provide adequate support for hard DLC films, adversely affecting their tribological performance and durability. An innovative concept in the development of advanced tribological systems involves coatings with a functionally-graded interface (FGI). This concept was experimentally studied in the present work by developing a model system involving DLC coatings with FGI based on theoretical finite element analysis (FEA) predictions. It was found that the durability of the DLC film is affected by the presence of the FGI and the loading level. Under low loading, FGI has a small effect due to the limited yielding occurring in the substrate. Under high loading conditions, the presence of FGI produced significant improvements (;80% increase in coating lifetime). In view of the FEA predictions, these effects can be attributed to reduction of the plastic zone size and prevention of yielding at the coatingy substrate interface. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Duplex treatment; DLC; Functionally gradient materials; Tribology
1. Introduction Diamond-like carbon (DLC) coatings have high potential for developing advanced tribological systems. Tribological properties of DLC films have been reported w1–9x and demonstrate high wear resistance and low coefficients of friction ( f -0.15). To date, the wear behavior of DLC films has been profoundly studied w10–13x. It should also be noted that most of these studies have been conducted on DLC films deposited on ‘hard’ substrates (ceramics and hardened steels), whereas very little work has been done on the properties of DLC on ‘soft’ substrates. Stress field analysis on sliding contacts under loading has established that when the coefficient of friction is reduced (as is the case with DLC films), significant shear stresses are developed and the location of the maximum shear stress moves gradually into the substrate away from the substrateycoating interface * Corresponding author. Tel.: q1 225 578 5806; fax: q1 225 578 5924. E-mail address: [email protected] (E.I. Meletis).
w14,15x. There is initial evidence w16x suggesting that under these conditions, soft substrate materials, i.e. Ti– 6Al–4V alloy, may not be able to provide adequate support for the hard DLC films, adversely affecting their tribological performance and durability. An innovative approach in improving tribological behavior in such cases is to design and develop systems incorporating multilayers andyor duplex diffusionycoating treatments w2,3,16,17x. Hardening (strengthening) of the substrate surface layer can inhibit deformation, reduce the abrupt change in properties at the substratey coating interface, thus minimizing effects from incompatible mechanical response. These factors can have a profound effect on film wear behavior and their understanding is needed for the development of advanced tribological systems. Recently, a finite element (FE) modeling effort was conducted to describe the initial yielding behavior of a system composed by a hard coatingygraded substrate. The results from this study have provided the theoretical background for designing the experimental system that was investigated in the present work w18x.
0257-8972/02/$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 7 - 8 9 7 2 Ž 0 1 . 0 1 6 8 8 - 7
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
Fig. 1. Profiles of calculated equivalent stress and yield strength of graded and ungraded substrate: s9, von Mises equivalent stress; Po, loading level; z, depth from the surface; a, contact radius; t, coating thickness; R, radius of indenter; f , coefficient of friction (tyRs0, PoyHDLCs0.185, f s0).
Alloy Ti–6Al–4V was selected as the substrate material for the development of the duplex treatments. Titanium and its alloys possess excellent strength-toweight ratio, fracture toughness, corrosion resistance and biocompatibility, but their wear performance is not satisfactory. Thus, there is a need to improve the wear properties of titanium alloys in order to broaden the field of applications of these alloys. Intensified plasmaassisted nitriding was used as a precursor treatment to harden the substrate and produce a functionally-graded interface (FGI), followed by ion-beam deposition of DLC film. 2. Summary of FE modeling on initial yielding behavior of hard coatingygraded substrate In a recent study, FEM was used to investigate the initial yielding behavior in a system composed of a hard coating deposited on a soft substrate with a graded surface layer w18x. As noted in the previous section, plasma nitriding of Ti alloys produces a nitrogen diffusion layer that results in a hardness gradient. This concept is demonstrated in Fig. 1 with the profiles of equivalent stress and yield strength of the graded substrate in the case of indentation with no coating present (tyRs0, tscoating thickness and Rsindentor radius). The results show that for a given loading level, the substrate with a proper gradient in yield strength can remain above the yield point at all depths while the original substrate is beyond the yield point. Furthermore, upon increasing loading level, the location of the initial yield point can be shifted significantly into the substrate (zo yas1.3 for the graded and 0.67 for the ungraded substrate; zosdepth into the substrate where initial yielding occurs and ascontact radius).
179
Also, the results indicated that coating thickness, contact stress and the ratio of coatingysubstrate properties affect the location of initial yielding point and the size of yielding zone in the coatingysubstrate system, as shown in Fig. 2. For the ungraded substrate with thin coatings (tyRs0.02, 0.04), yielding initiates in the substrate just beneath the contacting region (0.40a and 0.22a). For thicker coatings (tyR)0.07), yielding starts at the coatingysubstrate interface regardless of loading level. For the graded substrate, yielding occurs at the knee of the graded profile (1.3a) independent of coating thickness. Thus, a graded substrate offers several advantages and has the potential to improve the tribological behavior of the hard coating. These findings constituted the theoretical base for designing the experimental system that was studied in the present work. 3. Experimental design and procedures 3.1. Development of nitrided layer and deposition of DLC coatings Ion nitriding was conducted using intensified plasmaassisted processing (IPAP) that utilizes a triode glow discharge system w19x. Compared to conventional plasma nitriding, the arrangement in IPAP can achieve higher bombarding particle (ions and neutrals) energies leading to several beneficial surface effects. Prior to nitriding, the specimen surface was sputter cleaned by Arq at 2000 V accelerating voltage and 6.65 Pa partial pressure for 10 min. The nitriding was carried out at 2000 V cathode voltage, anode bias 80 V, cathode current density 1 mA cmy2 and partial pressure of nitrogen 6.65 Pa. A kinetic study was conducted to develop the relationship between processing time and depth of nitriding zone. Thus, nitriding time was selected based on required nitriding depth. Following nitriding,
Fig. 2. Location of initial yielding position as a function of coating thickness for graded and ungraded substrate: z0, location of initial yielding ( f s0.1).
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
180
carbon source) was fed into the high vacuum chamber and r.f. energy is utilized to break the bonding in the CH4 molecule. Also, a bias is applied to the substrate to attract the generated ions resulting in deposition of DLC film. A kinetic study on DLC produced a growth rate of ;70 nm miny1. High purity methane was fed into the chamber and its partial pressure was kept at 2.9 Pa throughout the deposition process. The r.f. forward power density during deposition was approximately 12 W cmy2 and the bias on the substrates was 500 V. The thickness of the DLC coatings in this study was approximately 0.4 mm. 3.2. Growth kinetics of nitrided layer The microhardness (Vicker’s indenter loaded at 0.5 N) profiles of Ti–6Al–4V specimens processed by intensified plasma nitriding (IPAP) for 4 and 9 h are shown in Fig. 3a. These specimens were processed at 6.65 Pa N2 pressure, 2 kV bias voltage and 1 mA cmy2 current density. These conditions were selected since they produce a nitrided surface with relatively low roughness. The nitriding depths of the specimens processed for 4 and 9 h were found to be 35 and 55 mm, respectively. Fig. 3b presents the growth kinetics of the nitriding depth for the aforementioned processing conditions. As expected, a linear relationship exists between the growth of the nitrogen diffusion layer and the square root of time showing a volume diffusion-controlled process. Similar observations have been made previously for intensified and conventional plasma nitriding w20x with the difference that the intensified process exhibits much faster kinetics w19x. 3.3. Yielding in DLC-coated Ti–6Al–4V alloy under indentation Fig. 3. (a) Hardness profiles of nitrided Ti–6Al–4V alloy and (b) growth kinetics of nitrided zone in Ti–6Al–4V alloy. Table 1 The values of contact radius and maximum contact stress between Al2O3 and nitridedyDLC Ti–6Al–4V alloy External loading (N)
Contact radius a (mm)
Max. contact stress P0 (GPa)
1 2.5 5
0.74 1.00 1.26
879 1194 1504
a thin Si bond layer (approx. 50 nm thick) was deposited by using a magnetron sputtering source. The purpose of the bond layer was to enhance adhesion between the substrate and the DLC w4,16x. A magnetron sputtering system (Perkin-Elmer, Model 2400) was utilized for conducting the deposition of DLC films. In this deposition system, methane (the
Pin-on-disc friction tests were conducted to investigate the tribological behavior of duplex DLC coatingy nitriding treatments. In the friction tests, Al2O3 was selected as the pin material with a diameter of 9.5 mm (Es343 GPa, ns0.25). Three loading levels, 1, 2.5 and 5 N, were applied which given the thickness for the DLC coating (;0.4 mm), produce an initial yielding point that lies within the substrate w18x. The contact between a ball and a flat surface is in the axisymmetric condition. A theoretical analysis was performed on the complete stress field due to a circular contact region carrying a ‘hemisphere’ Hertzian normal pressure and proportionally distributed shearing traction w21x. It was assumed that during indentation of a sphere on a flat surface, the distribution of normal contact stress, P(r), on the contact area is in the elliptical form within the contact width, 2a PŽr.sP0
y
1y
r2 a2
Žfor)r)Fa.,
(1)
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
181
4a and the real values of von Mises equivalent stress in DLC-coated Ti–6Al–4V alloy are plotted in Fig. 4b for external loading levels of 1, 2.5 and 5 N. The intercepts of the plots with the stress value 1.04 GPa (yield strength of Ti–6Al–4V alloy), Dzya, are 6.2, 7.6 and 9.6 corresponding to loading levels of 1, 2.5 and 5 N, respectively, which gives the approximate size of the plastic zone under those loading levels, Dzs4.57, 7.60 and 12.1 mm, respectively. It should be noted that the above analysis is based on the assumption of uniform material (without coating) that is reasonable in view of the small thickness of the coating. According to the FE modeling, in the case of indentation in a hard coatingysoft substrate system, due to the higher load carrying capacity of the hard coating, the size of plastic zone within the substrate can be decreased compared with the case without coating w18x. Thus, the estimated sizes of plastic zone for the indentation of spherical Al2O3 yDLC coated Ti–6Al–4V alloy are the upper limits of the plastic zone in the substrate under various loading levels. 3.4. Determination of the nitriding depth
Fig. 4. (a) Normalized svm distribution along the depth and (b) svm distribution for the three loading levels used in the present study.
3P (P is the external load) is the maxi2pa2 mum contact stress at the center of the contact region, and a is the contact radius. Assuming contact of the Al2O3 ball with the flat DLC-coated surface and neglecting the influence of the substrate properties, the calculated values of P0 and a under 1, 2.5 and 5 N are listed in Table 1. Under the boundary condition given by Eq. (1) and ignoring the effect of tangential traction due to low friction of DLC coating, the normalized tensile stress along the centerline of contact region (z-axis) can be expressed by w21x where P0s
w B a Ez sxx syy z s sŽ1qn.x1y tany1C F| D z G~ P0 P0 a y 2 y1 1B z E q C1q 2 F 2D a G
szz B z2 Ey1 sC1q 2 F P0 D a G
(2a)
The plastic deformation of the surface region in the substrate during loading should be prevented in order to accommodate the hard DLC coating. Based on the size of plastic zone under the present loading levels estimated in the previous section, the depth of nitrogen diffusion zone needs to be G12 mm. The optimum operational parameters for the nitrogen diffusion process by IPAP have been obtained from the results of the preliminary experimental study. Two nitriding depths of ;15 and ;30 mm were developed in the present study to investigate the effect of the depth of the hardened zone on the tribological behavior of DLC films. According to the kinetic study (Fig. 3b), these two values of nitriding depth approximately correspond to nitriding times of approximately 1 and 4 h, respectively. All friction testing on nitridedyDLC Ti–6Al–4V alloy was conducted at a sliding velocity of 0.3 m sy1, at room temperature (;25 8C) and under room humidity (;40%). The total sliding distance for all tests was 10 km or until the DLC coatings failed. Wear rates of disc (DLC film) and pin materials were determined by profilometry and optical microscopy, respectively, as has been previously described w10–13x. 4. Experimental results and discussion
(2b)
where sxx, syy and szz is the von Misses equivalent stress in the x, y, z direction, respectively, and n is Poisson’s ratio. The distribution of normalized von Mises equivalent stress derived from Eqs. (2a) and (2b) is shown in Fig.
The measured hardness of as-deposited DLC coating was approximately 26 GPa and is very close to the hardness value of DLC films produced by IBD w10x. Friction tests were conducted for three different treatments: (a) DLC coatingyTi–6Al–4V alloy (unprocessed); (b) DLC coatingyTi–6Al–4V alloy with 15-mm
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
182
Table 2 Friction tests of Al2O3yDLC-coated Ti–6Al–4V and nitrided Ti6Al–4V alloy Test sample (Al2O3 pin)
Test no
Load (N)
sH (MPa)
f in
f ss
Sf (km)
WB9 (=10y8 mm3 my1 Ny1)
WD9 (=10y7 mm3 my1 Ny1)
(a) DLCyTi–6Al–4V
1 2 3 4 5 6 7 8 9
1 2.5 5 1 2.5 5 1 2.5 5
513 697 878 513 697 878 513 697 878
0.14 0.17 0.14 0.04 0.13 0.10 0.10 0.13 0.12
0.20 0.15 0.11 0.08 0.17 0.12 0.15 0.18 0.15
)10 6.12 3.67 )10 6.44 5.96 )10 8.91 6.58
UD 0.173 0.187 UD 0.366 0.843 UD 0.375 1.07
2.28 3.54 2.75 2.75 3.37 3.09 2.29 3.13 2.34
(b) NitridedyDLC Ti–6Al–4V (15 mm) (c) NitridedyDLC Ti–6Al–4V (30 mm)
sH, initial Hertzian stress; f in and f ss, initial and steady-state coefficient of friction, respectively; Sf , sliding distance before failure of DLC coating; WB9 and WD9, wear rate of ball and disc, respectively; UD, undetectable.
nitrided zone; (c) DLC coating yTi–6Al–4V alloy with 30 mm nitrided zone. The results from the friction tests are summarized in Table 2. In this study, the durability of the DLC film is defined as the sliding distance, Sf, required for the DLC film to fail. Several interesting observations can be made by analyzing the results shown in Table 2. For example, in the samples with the same treatment, the wear rate of the pin increases by increasing loading level. This can be attributed to the larger contact area between pin and disc due to higher external loading. Also, the wear rate of the pin depends on the hardness of the counter-face, i.e. DLCysubstrate system. The comparison of the wear rate of the Al2O3 pin among tests 2, 5 and 8 (0.173=10y8, 0.366=10y8 and 0.375=10y8 mm3 my1 Ny1, respectively), and among tests 3, 6 and 9 (0.187=10y8, 0.843=10y8 and 1.07=10y8 mm3 my1 Ny1, respectively) indicates that the substrate hardening increases the overall hardness of the counter-face w16x that is a result the progressive increase of the pin wear rate. In a system consisting of a thin DLC coating deposited on a nitriding-hardened surface, the strength response of the layered system to the indentation by a high load is influenced by the properties of the substrate w22x. The degree of the influence depends on the ratio of the coating thickness and indentation depth. For the same coating thickness, higher loading level results in a larger indentation depth and thus increases the effect of substrate properties on the friction behavior. This effect can be seen by comparing pin wear rates at the same loading for the various treatments (increasing substrate hardness). As revealed by the data in Table 2, for samples within the same treatment, increasing loading level decreased the lifetime (Sf) of the DLC coating. It has been found that during friction the primary cause of failure of DLC films is delamination from the substrate w23x. According to the previous FE analysis w18x, when external loading is applied to a thin filmygraded substrate system, yielding initiates at a certain depth in the substrate, and
gradually grows in both directions, i.e. into the substrate and towards the coatingysubstrate interface. When the yielding zone reaches the interface, it is expected to create interfacial deflection of the substrate. Because of the mismatch of the deflection between the hard DLC film and soft substrate during indentation and friction, microcracks can nucleate at the DLCysubstrate interface causing film delamination. Furthermore, based on the profiles of the elasticyplastic boundaries given by Stephens et al. w18x, it can be seen that a higher external contact stress increases the plastic region at the interface. Thus, under higher external loading, microcrack nucleation at the interface is easier reducing film durability. Another interesting observation was that in all tests, the wear rate of the DLC film did not exhibit any significant variation (wear rates were determined based on film durability, Sf, and varied between 2–3=10y7 mm3 my1 Ny1). This observation suggests that surface treatment of the substrate does not exercise a significant influence on the wear rate of the DLC coating. Since the sputtered Si interlayer can form strong chemical bonding with DLC structure and thus enhance the adhesion between DLC film and its substrate, the wear rate of DLC coating depends mainly on its as-deposited structure and residual stress, as proposed in a previous study w23x. However, the surface treatment has a major effect on the failure mechanism as is discussed later in this section. Tests 1, 4 and 7 present the tribological behavior of DLCynitrided substrate system under low loading level (1 N). The results showed that for the present total sliding distance (10 km) the durability of the DLC films is not affected by the substrate surface treatment. Additionally, the wear of the Al2O3 pin was negligible. Based on the information provided in by Stephens et al. w18x, it is evident that under a low external contact stress the substrate plastic deformation is limited (if any). These conditions allow the interface region to operate mainly in the elastic regime and no effect from substrate hardening can be realized. Thus, in this case the lifetime
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
of the film is a result of the gradual wearing out of the coating by the hard Al2O3 pin. Under medium loading (2.5 N), DLC film durability was found to increase with substrate hardening (6.12, 6.44 and 8.91 km, for unprocessed, 15 and 30 mm nitrided zone, respectively). The observed increase in durability indicates that the modified substrate has a positive impact on the tribological behavior of the DLC coatings. Based on the FE analysis, substrate hardening has two major effects. It can effectively reduce (or even eliminate) the plastic zone size and restrict its growth by not allowing it to reach the coatingysubstrate interface (Fig. 2). This is evident by the observed 5.2 and 45.6% higher durability produced by the 15 and 30-mm nitrided zone, respectively. Tests 3, 6 and 9 exhibit the tribological behavior of DLCynitrided substrate system under high loading level (5 N). Compared with the durability of DLCyuntreated substrate (3.67 km), development of a nitrided zone produced significant improvements (5.96 and 6.58 km, for 15 and 30-mm nitrided depth, corresponding to 62.4 and 79.3% improvement, respectively). It is evident from these results that increasing the loading level makes the effect of the substrate hardening more pronounced. In brief, the lifetime of DLC with FGI produced by nitriding process is significantly improved, compared with that of DLCyunprocessed substrate. More enhancement on the durability of DLC coating is found under high loading level. This experimental response is consistent with the FE analysis in the previous report w18x. 5. Conclusions Substrate hardening by intensified plasma nitriding can produce a FGI that was found to positively impact on the durability of DLC coatings. At low external loading, the FGI exercised a small effect on durability of DLC because the interfacial region of the coatingysubstrate system is mainly in the elastic regime. In this case the lifetime of the DLC coating is determined by the wear process that is gradually consuming the film. At higher loading levels, the presence of the FGI resulted in a progressively more significant effect that reaches an improvement of approximately 80% in durability for the loading levels applied in the present study. These positive effects are attributed to the reduction of the plastic zone size developed in the substrate and the prevention of yielding from reaching the coatingysubstrate interface. Thus, substrate hardening by intensified
183
plasma nitriding provides the means for the development of an effective FGI that can affect the yielding behavior of the substrate. Acknowledgments This work was supported by the Army Research Office (Grant DAAG55-98-1-0279) and Louisiana Board of Regents. References w1x M.S. Wong, R. Meilunas, T.P. Ong, R.P. Chang, in: L.E. Pope, L.L. Fehrenbacher, W.O. Winer (Eds.), New Materials Approaches to Tribology: Theory and Applications, MRS, Pittsburgh, Pennsylvania, 1991, p. 483. w2x A. Imamura, T. Tsukamoto, K. Shibuki, S. Takatsu, Surf. Coat. Technol. 36 (1988) 161. w3x K. Enke, H. Dimigen, H. Hubsch, Appl. Phys. Lett. 36 (1980) 291. w4x A. Erdemir, M. Switala, R. Wei, P. Wilbur, Surf. Coat. Technol. 50 (1991) 17. w5x K. Miyoshi, R.L. Wu, A. Garscadden, Surf. Coat. Technol. 54 y55 (1992) 428. w6x S.D. Gorpinchenko, S.M. Klotsman, E.V. Kuzmina, Surf. Coat. Technol. 47 (1991) 201. w7x A. Grill, V. Patel, B.S. Meyerson, Surf. Coat. Technol. 49 (1991) 530. w8x D.S. Kim, T.E. Fisher, B. Gallois, Surf. Coat. Technol. 49 (1991) 537. w9x A. Erdemir, F.A. Nichols, X.Z. Pan, R. Wei, P. Wilbur, Diamond Relat. Mater. 3 (1993) 119. w10x Y. Liu, A. Erdemir, E.I. Meletis, Surf. Coat. Technol. 82 (1996) 48. w11x Y. Liu, E.I. Meletis, J. Mater. Sci. 32 (1997) 3491. w12x Y. Liu, A. Erdemir, E.I. Meletis, Surf. Coat. Technol. 86-87 (1996) 564. w13x Y. Liu, A. Erdemir, E.I. Meletis, Surf. Coat. Technol. 94-95 (1997) 463. w14x G.M. Hamilton, L.E. Goodman, J. Appl. Mech. 33 (1966) 371. w15x D. Diao, K. Kato, Thin Solid Films 245 (1994) 115. w16x E.I. Meletis, A. Erdemir, G.R. Fenske, Surf. Coat. Technol. 73 (1995) 39. w17x A.A. Voevedin, C. Rebholz, J.M. Schneider, P. Stevenson, A. Matthews, Surf. Coat. Technol. 73 (1995) 47. w18x L.S. Stephens, Y. Liu, E.I. Meletis, J. Tribol. Trans. ASME 122 (2000) 381. w19x A. Adjaottor, E. Ma, E.I. Meletis, Surf. Coat. Technol. 89 (1997) 197. w20x K.T. Rie, T. Lampe, Mater. Sci. Eng. 69 (1985) 473. w21x G.M. Hamilton, L.E. Goodman, J. Appl. Mech. Trans. ASME 33 (1966) 371. w22x Y. Sun, T. Bell, S. Zheng, Thin Solid Films 258 (1995) 198. w23x R. Wei, P.J. Wilbur, A. Erdemir, F.M. Kustas, Surf. Coat. Technol. 51 (1992) 139.
Tribological behavior of DLC coatings with functionally gradient interfaces Y. Liu, E.I. Meletis* Materials Science and Engineering Program, Mechanical Engineering Department, Louisiana State University, Baton Rouge, LA 70803, USA Received 6 June 2001; accepted in revised form 14 November 2001
Abstract Initial evidence showed that soft substrate materials (such as Ti–6Al–4V alloy) might not be able to provide adequate support for hard DLC films, adversely affecting their tribological performance and durability. An innovative concept in the development of advanced tribological systems involves coatings with a functionally-graded interface (FGI). This concept was experimentally studied in the present work by developing a model system involving DLC coatings with FGI based on theoretical finite element analysis (FEA) predictions. It was found that the durability of the DLC film is affected by the presence of the FGI and the loading level. Under low loading, FGI has a small effect due to the limited yielding occurring in the substrate. Under high loading conditions, the presence of FGI produced significant improvements (;80% increase in coating lifetime). In view of the FEA predictions, these effects can be attributed to reduction of the plastic zone size and prevention of yielding at the coatingy substrate interface. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Duplex treatment; DLC; Functionally gradient materials; Tribology
1. Introduction Diamond-like carbon (DLC) coatings have high potential for developing advanced tribological systems. Tribological properties of DLC films have been reported w1–9x and demonstrate high wear resistance and low coefficients of friction ( f -0.15). To date, the wear behavior of DLC films has been profoundly studied w10–13x. It should also be noted that most of these studies have been conducted on DLC films deposited on ‘hard’ substrates (ceramics and hardened steels), whereas very little work has been done on the properties of DLC on ‘soft’ substrates. Stress field analysis on sliding contacts under loading has established that when the coefficient of friction is reduced (as is the case with DLC films), significant shear stresses are developed and the location of the maximum shear stress moves gradually into the substrate away from the substrateycoating interface * Corresponding author. Tel.: q1 225 578 5806; fax: q1 225 578 5924. E-mail address: [email protected] (E.I. Meletis).
w14,15x. There is initial evidence w16x suggesting that under these conditions, soft substrate materials, i.e. Ti– 6Al–4V alloy, may not be able to provide adequate support for the hard DLC films, adversely affecting their tribological performance and durability. An innovative approach in improving tribological behavior in such cases is to design and develop systems incorporating multilayers andyor duplex diffusionycoating treatments w2,3,16,17x. Hardening (strengthening) of the substrate surface layer can inhibit deformation, reduce the abrupt change in properties at the substratey coating interface, thus minimizing effects from incompatible mechanical response. These factors can have a profound effect on film wear behavior and their understanding is needed for the development of advanced tribological systems. Recently, a finite element (FE) modeling effort was conducted to describe the initial yielding behavior of a system composed by a hard coatingygraded substrate. The results from this study have provided the theoretical background for designing the experimental system that was investigated in the present work w18x.
0257-8972/02/$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 7 - 8 9 7 2 Ž 0 1 . 0 1 6 8 8 - 7
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
Fig. 1. Profiles of calculated equivalent stress and yield strength of graded and ungraded substrate: s9, von Mises equivalent stress; Po, loading level; z, depth from the surface; a, contact radius; t, coating thickness; R, radius of indenter; f , coefficient of friction (tyRs0, PoyHDLCs0.185, f s0).
Alloy Ti–6Al–4V was selected as the substrate material for the development of the duplex treatments. Titanium and its alloys possess excellent strength-toweight ratio, fracture toughness, corrosion resistance and biocompatibility, but their wear performance is not satisfactory. Thus, there is a need to improve the wear properties of titanium alloys in order to broaden the field of applications of these alloys. Intensified plasmaassisted nitriding was used as a precursor treatment to harden the substrate and produce a functionally-graded interface (FGI), followed by ion-beam deposition of DLC film. 2. Summary of FE modeling on initial yielding behavior of hard coatingygraded substrate In a recent study, FEM was used to investigate the initial yielding behavior in a system composed of a hard coating deposited on a soft substrate with a graded surface layer w18x. As noted in the previous section, plasma nitriding of Ti alloys produces a nitrogen diffusion layer that results in a hardness gradient. This concept is demonstrated in Fig. 1 with the profiles of equivalent stress and yield strength of the graded substrate in the case of indentation with no coating present (tyRs0, tscoating thickness and Rsindentor radius). The results show that for a given loading level, the substrate with a proper gradient in yield strength can remain above the yield point at all depths while the original substrate is beyond the yield point. Furthermore, upon increasing loading level, the location of the initial yield point can be shifted significantly into the substrate (zo yas1.3 for the graded and 0.67 for the ungraded substrate; zosdepth into the substrate where initial yielding occurs and ascontact radius).
179
Also, the results indicated that coating thickness, contact stress and the ratio of coatingysubstrate properties affect the location of initial yielding point and the size of yielding zone in the coatingysubstrate system, as shown in Fig. 2. For the ungraded substrate with thin coatings (tyRs0.02, 0.04), yielding initiates in the substrate just beneath the contacting region (0.40a and 0.22a). For thicker coatings (tyR)0.07), yielding starts at the coatingysubstrate interface regardless of loading level. For the graded substrate, yielding occurs at the knee of the graded profile (1.3a) independent of coating thickness. Thus, a graded substrate offers several advantages and has the potential to improve the tribological behavior of the hard coating. These findings constituted the theoretical base for designing the experimental system that was studied in the present work. 3. Experimental design and procedures 3.1. Development of nitrided layer and deposition of DLC coatings Ion nitriding was conducted using intensified plasmaassisted processing (IPAP) that utilizes a triode glow discharge system w19x. Compared to conventional plasma nitriding, the arrangement in IPAP can achieve higher bombarding particle (ions and neutrals) energies leading to several beneficial surface effects. Prior to nitriding, the specimen surface was sputter cleaned by Arq at 2000 V accelerating voltage and 6.65 Pa partial pressure for 10 min. The nitriding was carried out at 2000 V cathode voltage, anode bias 80 V, cathode current density 1 mA cmy2 and partial pressure of nitrogen 6.65 Pa. A kinetic study was conducted to develop the relationship between processing time and depth of nitriding zone. Thus, nitriding time was selected based on required nitriding depth. Following nitriding,
Fig. 2. Location of initial yielding position as a function of coating thickness for graded and ungraded substrate: z0, location of initial yielding ( f s0.1).
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
180
carbon source) was fed into the high vacuum chamber and r.f. energy is utilized to break the bonding in the CH4 molecule. Also, a bias is applied to the substrate to attract the generated ions resulting in deposition of DLC film. A kinetic study on DLC produced a growth rate of ;70 nm miny1. High purity methane was fed into the chamber and its partial pressure was kept at 2.9 Pa throughout the deposition process. The r.f. forward power density during deposition was approximately 12 W cmy2 and the bias on the substrates was 500 V. The thickness of the DLC coatings in this study was approximately 0.4 mm. 3.2. Growth kinetics of nitrided layer The microhardness (Vicker’s indenter loaded at 0.5 N) profiles of Ti–6Al–4V specimens processed by intensified plasma nitriding (IPAP) for 4 and 9 h are shown in Fig. 3a. These specimens were processed at 6.65 Pa N2 pressure, 2 kV bias voltage and 1 mA cmy2 current density. These conditions were selected since they produce a nitrided surface with relatively low roughness. The nitriding depths of the specimens processed for 4 and 9 h were found to be 35 and 55 mm, respectively. Fig. 3b presents the growth kinetics of the nitriding depth for the aforementioned processing conditions. As expected, a linear relationship exists between the growth of the nitrogen diffusion layer and the square root of time showing a volume diffusion-controlled process. Similar observations have been made previously for intensified and conventional plasma nitriding w20x with the difference that the intensified process exhibits much faster kinetics w19x. 3.3. Yielding in DLC-coated Ti–6Al–4V alloy under indentation Fig. 3. (a) Hardness profiles of nitrided Ti–6Al–4V alloy and (b) growth kinetics of nitrided zone in Ti–6Al–4V alloy. Table 1 The values of contact radius and maximum contact stress between Al2O3 and nitridedyDLC Ti–6Al–4V alloy External loading (N)
Contact radius a (mm)
Max. contact stress P0 (GPa)
1 2.5 5
0.74 1.00 1.26
879 1194 1504
a thin Si bond layer (approx. 50 nm thick) was deposited by using a magnetron sputtering source. The purpose of the bond layer was to enhance adhesion between the substrate and the DLC w4,16x. A magnetron sputtering system (Perkin-Elmer, Model 2400) was utilized for conducting the deposition of DLC films. In this deposition system, methane (the
Pin-on-disc friction tests were conducted to investigate the tribological behavior of duplex DLC coatingy nitriding treatments. In the friction tests, Al2O3 was selected as the pin material with a diameter of 9.5 mm (Es343 GPa, ns0.25). Three loading levels, 1, 2.5 and 5 N, were applied which given the thickness for the DLC coating (;0.4 mm), produce an initial yielding point that lies within the substrate w18x. The contact between a ball and a flat surface is in the axisymmetric condition. A theoretical analysis was performed on the complete stress field due to a circular contact region carrying a ‘hemisphere’ Hertzian normal pressure and proportionally distributed shearing traction w21x. It was assumed that during indentation of a sphere on a flat surface, the distribution of normal contact stress, P(r), on the contact area is in the elliptical form within the contact width, 2a PŽr.sP0
y
1y
r2 a2
Žfor)r)Fa.,
(1)
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
181
4a and the real values of von Mises equivalent stress in DLC-coated Ti–6Al–4V alloy are plotted in Fig. 4b for external loading levels of 1, 2.5 and 5 N. The intercepts of the plots with the stress value 1.04 GPa (yield strength of Ti–6Al–4V alloy), Dzya, are 6.2, 7.6 and 9.6 corresponding to loading levels of 1, 2.5 and 5 N, respectively, which gives the approximate size of the plastic zone under those loading levels, Dzs4.57, 7.60 and 12.1 mm, respectively. It should be noted that the above analysis is based on the assumption of uniform material (without coating) that is reasonable in view of the small thickness of the coating. According to the FE modeling, in the case of indentation in a hard coatingysoft substrate system, due to the higher load carrying capacity of the hard coating, the size of plastic zone within the substrate can be decreased compared with the case without coating w18x. Thus, the estimated sizes of plastic zone for the indentation of spherical Al2O3 yDLC coated Ti–6Al–4V alloy are the upper limits of the plastic zone in the substrate under various loading levels. 3.4. Determination of the nitriding depth
Fig. 4. (a) Normalized svm distribution along the depth and (b) svm distribution for the three loading levels used in the present study.
3P (P is the external load) is the maxi2pa2 mum contact stress at the center of the contact region, and a is the contact radius. Assuming contact of the Al2O3 ball with the flat DLC-coated surface and neglecting the influence of the substrate properties, the calculated values of P0 and a under 1, 2.5 and 5 N are listed in Table 1. Under the boundary condition given by Eq. (1) and ignoring the effect of tangential traction due to low friction of DLC coating, the normalized tensile stress along the centerline of contact region (z-axis) can be expressed by w21x where P0s
w B a Ez sxx syy z s sŽ1qn.x1y tany1C F| D z G~ P0 P0 a y 2 y1 1B z E q C1q 2 F 2D a G
szz B z2 Ey1 sC1q 2 F P0 D a G
(2a)
The plastic deformation of the surface region in the substrate during loading should be prevented in order to accommodate the hard DLC coating. Based on the size of plastic zone under the present loading levels estimated in the previous section, the depth of nitrogen diffusion zone needs to be G12 mm. The optimum operational parameters for the nitrogen diffusion process by IPAP have been obtained from the results of the preliminary experimental study. Two nitriding depths of ;15 and ;30 mm were developed in the present study to investigate the effect of the depth of the hardened zone on the tribological behavior of DLC films. According to the kinetic study (Fig. 3b), these two values of nitriding depth approximately correspond to nitriding times of approximately 1 and 4 h, respectively. All friction testing on nitridedyDLC Ti–6Al–4V alloy was conducted at a sliding velocity of 0.3 m sy1, at room temperature (;25 8C) and under room humidity (;40%). The total sliding distance for all tests was 10 km or until the DLC coatings failed. Wear rates of disc (DLC film) and pin materials were determined by profilometry and optical microscopy, respectively, as has been previously described w10–13x. 4. Experimental results and discussion
(2b)
where sxx, syy and szz is the von Misses equivalent stress in the x, y, z direction, respectively, and n is Poisson’s ratio. The distribution of normalized von Mises equivalent stress derived from Eqs. (2a) and (2b) is shown in Fig.
The measured hardness of as-deposited DLC coating was approximately 26 GPa and is very close to the hardness value of DLC films produced by IBD w10x. Friction tests were conducted for three different treatments: (a) DLC coatingyTi–6Al–4V alloy (unprocessed); (b) DLC coatingyTi–6Al–4V alloy with 15-mm
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
182
Table 2 Friction tests of Al2O3yDLC-coated Ti–6Al–4V and nitrided Ti6Al–4V alloy Test sample (Al2O3 pin)
Test no
Load (N)
sH (MPa)
f in
f ss
Sf (km)
WB9 (=10y8 mm3 my1 Ny1)
WD9 (=10y7 mm3 my1 Ny1)
(a) DLCyTi–6Al–4V
1 2 3 4 5 6 7 8 9
1 2.5 5 1 2.5 5 1 2.5 5
513 697 878 513 697 878 513 697 878
0.14 0.17 0.14 0.04 0.13 0.10 0.10 0.13 0.12
0.20 0.15 0.11 0.08 0.17 0.12 0.15 0.18 0.15
)10 6.12 3.67 )10 6.44 5.96 )10 8.91 6.58
UD 0.173 0.187 UD 0.366 0.843 UD 0.375 1.07
2.28 3.54 2.75 2.75 3.37 3.09 2.29 3.13 2.34
(b) NitridedyDLC Ti–6Al–4V (15 mm) (c) NitridedyDLC Ti–6Al–4V (30 mm)
sH, initial Hertzian stress; f in and f ss, initial and steady-state coefficient of friction, respectively; Sf , sliding distance before failure of DLC coating; WB9 and WD9, wear rate of ball and disc, respectively; UD, undetectable.
nitrided zone; (c) DLC coating yTi–6Al–4V alloy with 30 mm nitrided zone. The results from the friction tests are summarized in Table 2. In this study, the durability of the DLC film is defined as the sliding distance, Sf, required for the DLC film to fail. Several interesting observations can be made by analyzing the results shown in Table 2. For example, in the samples with the same treatment, the wear rate of the pin increases by increasing loading level. This can be attributed to the larger contact area between pin and disc due to higher external loading. Also, the wear rate of the pin depends on the hardness of the counter-face, i.e. DLCysubstrate system. The comparison of the wear rate of the Al2O3 pin among tests 2, 5 and 8 (0.173=10y8, 0.366=10y8 and 0.375=10y8 mm3 my1 Ny1, respectively), and among tests 3, 6 and 9 (0.187=10y8, 0.843=10y8 and 1.07=10y8 mm3 my1 Ny1, respectively) indicates that the substrate hardening increases the overall hardness of the counter-face w16x that is a result the progressive increase of the pin wear rate. In a system consisting of a thin DLC coating deposited on a nitriding-hardened surface, the strength response of the layered system to the indentation by a high load is influenced by the properties of the substrate w22x. The degree of the influence depends on the ratio of the coating thickness and indentation depth. For the same coating thickness, higher loading level results in a larger indentation depth and thus increases the effect of substrate properties on the friction behavior. This effect can be seen by comparing pin wear rates at the same loading for the various treatments (increasing substrate hardness). As revealed by the data in Table 2, for samples within the same treatment, increasing loading level decreased the lifetime (Sf) of the DLC coating. It has been found that during friction the primary cause of failure of DLC films is delamination from the substrate w23x. According to the previous FE analysis w18x, when external loading is applied to a thin filmygraded substrate system, yielding initiates at a certain depth in the substrate, and
gradually grows in both directions, i.e. into the substrate and towards the coatingysubstrate interface. When the yielding zone reaches the interface, it is expected to create interfacial deflection of the substrate. Because of the mismatch of the deflection between the hard DLC film and soft substrate during indentation and friction, microcracks can nucleate at the DLCysubstrate interface causing film delamination. Furthermore, based on the profiles of the elasticyplastic boundaries given by Stephens et al. w18x, it can be seen that a higher external contact stress increases the plastic region at the interface. Thus, under higher external loading, microcrack nucleation at the interface is easier reducing film durability. Another interesting observation was that in all tests, the wear rate of the DLC film did not exhibit any significant variation (wear rates were determined based on film durability, Sf, and varied between 2–3=10y7 mm3 my1 Ny1). This observation suggests that surface treatment of the substrate does not exercise a significant influence on the wear rate of the DLC coating. Since the sputtered Si interlayer can form strong chemical bonding with DLC structure and thus enhance the adhesion between DLC film and its substrate, the wear rate of DLC coating depends mainly on its as-deposited structure and residual stress, as proposed in a previous study w23x. However, the surface treatment has a major effect on the failure mechanism as is discussed later in this section. Tests 1, 4 and 7 present the tribological behavior of DLCynitrided substrate system under low loading level (1 N). The results showed that for the present total sliding distance (10 km) the durability of the DLC films is not affected by the substrate surface treatment. Additionally, the wear of the Al2O3 pin was negligible. Based on the information provided in by Stephens et al. w18x, it is evident that under a low external contact stress the substrate plastic deformation is limited (if any). These conditions allow the interface region to operate mainly in the elastic regime and no effect from substrate hardening can be realized. Thus, in this case the lifetime
Y. Liu, E.I. Meletis / Surface and Coatings Technology 153 (2002) 178–183
of the film is a result of the gradual wearing out of the coating by the hard Al2O3 pin. Under medium loading (2.5 N), DLC film durability was found to increase with substrate hardening (6.12, 6.44 and 8.91 km, for unprocessed, 15 and 30 mm nitrided zone, respectively). The observed increase in durability indicates that the modified substrate has a positive impact on the tribological behavior of the DLC coatings. Based on the FE analysis, substrate hardening has two major effects. It can effectively reduce (or even eliminate) the plastic zone size and restrict its growth by not allowing it to reach the coatingysubstrate interface (Fig. 2). This is evident by the observed 5.2 and 45.6% higher durability produced by the 15 and 30-mm nitrided zone, respectively. Tests 3, 6 and 9 exhibit the tribological behavior of DLCynitrided substrate system under high loading level (5 N). Compared with the durability of DLCyuntreated substrate (3.67 km), development of a nitrided zone produced significant improvements (5.96 and 6.58 km, for 15 and 30-mm nitrided depth, corresponding to 62.4 and 79.3% improvement, respectively). It is evident from these results that increasing the loading level makes the effect of the substrate hardening more pronounced. In brief, the lifetime of DLC with FGI produced by nitriding process is significantly improved, compared with that of DLCyunprocessed substrate. More enhancement on the durability of DLC coating is found under high loading level. This experimental response is consistent with the FE analysis in the previous report w18x. 5. Conclusions Substrate hardening by intensified plasma nitriding can produce a FGI that was found to positively impact on the durability of DLC coatings. At low external loading, the FGI exercised a small effect on durability of DLC because the interfacial region of the coatingysubstrate system is mainly in the elastic regime. In this case the lifetime of the DLC coating is determined by the wear process that is gradually consuming the film. At higher loading levels, the presence of the FGI resulted in a progressively more significant effect that reaches an improvement of approximately 80% in durability for the loading levels applied in the present study. These positive effects are attributed to the reduction of the plastic zone size developed in the substrate and the prevention of yielding from reaching the coatingysubstrate interface. Thus, substrate hardening by intensified
183
plasma nitriding provides the means for the development of an effective FGI that can affect the yielding behavior of the substrate. Acknowledgments This work was supported by the Army Research Office (Grant DAAG55-98-1-0279) and Louisiana Board of Regents. References w1x M.S. Wong, R. Meilunas, T.P. Ong, R.P. Chang, in: L.E. Pope, L.L. Fehrenbacher, W.O. Winer (Eds.), New Materials Approaches to Tribology: Theory and Applications, MRS, Pittsburgh, Pennsylvania, 1991, p. 483. w2x A. Imamura, T. Tsukamoto, K. Shibuki, S. Takatsu, Surf. Coat. Technol. 36 (1988) 161. w3x K. Enke, H. Dimigen, H. Hubsch, Appl. Phys. Lett. 36 (1980) 291. w4x A. Erdemir, M. Switala, R. Wei, P. Wilbur, Surf. Coat. Technol. 50 (1991) 17. w5x K. Miyoshi, R.L. Wu, A. Garscadden, Surf. Coat. Technol. 54 y55 (1992) 428. w6x S.D. Gorpinchenko, S.M. Klotsman, E.V. Kuzmina, Surf. Coat. Technol. 47 (1991) 201. w7x A. Grill, V. Patel, B.S. Meyerson, Surf. Coat. Technol. 49 (1991) 530. w8x D.S. Kim, T.E. Fisher, B. Gallois, Surf. Coat. Technol. 49 (1991) 537. w9x A. Erdemir, F.A. Nichols, X.Z. Pan, R. Wei, P. Wilbur, Diamond Relat. Mater. 3 (1993) 119. w10x Y. Liu, A. Erdemir, E.I. Meletis, Surf. Coat. Technol. 82 (1996) 48. w11x Y. Liu, E.I. Meletis, J. Mater. Sci. 32 (1997) 3491. w12x Y. Liu, A. Erdemir, E.I. Meletis, Surf. Coat. Technol. 86-87 (1996) 564. w13x Y. Liu, A. Erdemir, E.I. Meletis, Surf. Coat. Technol. 94-95 (1997) 463. w14x G.M. Hamilton, L.E. Goodman, J. Appl. Mech. 33 (1966) 371. w15x D. Diao, K. Kato, Thin Solid Films 245 (1994) 115. w16x E.I. Meletis, A. Erdemir, G.R. Fenske, Surf. Coat. Technol. 73 (1995) 39. w17x A.A. Voevedin, C. Rebholz, J.M. Schneider, P. Stevenson, A. Matthews, Surf. Coat. Technol. 73 (1995) 47. w18x L.S. Stephens, Y. Liu, E.I. Meletis, J. Tribol. Trans. ASME 122 (2000) 381. w19x A. Adjaottor, E. Ma, E.I. Meletis, Surf. Coat. Technol. 89 (1997) 197. w20x K.T. Rie, T. Lampe, Mater. Sci. Eng. 69 (1985) 473. w21x G.M. Hamilton, L.E. Goodman, J. Appl. Mech. Trans. ASME 33 (1966) 371. w22x Y. Sun, T. Bell, S. Zheng, Thin Solid Films 258 (1995) 198. w23x R. Wei, P.J. Wilbur, A. Erdemir, F.M. Kustas, Surf. Coat. Technol. 51 (1992) 139.